Depending on break save states, decryption of certain slots sometimes never completes

[Kaisonic]

If the saves used to break encryption (save1, save2, save3, a.k.a. 26.sav, 16.sav, 165.sav) contain Pokemon in boxes other than 1 and 2, decryption of some slots fails, resulting in no Pokemon listed in a slot where there is Pokemon, and listing <100% Pokemon in slots that are empty.

For example, if I break encryption using three saves that contain ONLY the six Pokemon in the proper spots in the first two boxes, then I open saves that are empty, full1, and full2, decryption is fully successful. When I then re-open empty, no Pokemon appear, and when I re-open full1 or full2, all 930 Pokemon appear with no ~ markers. This is correct behavior.

However, if I break encryption using three saves that contain only the proper six Pokemon in the first two boxes, but the rest of the boxes contain random Pokemon, then I open the SAME three saves (empty, full1, full2), decryption is not fully successful. When I re-open empty, a single ~ marked Pokemon appears. When I re-open full1 or full2, only 929 Pokemon appear, and the slot that does not appear matches the slot that appeared in empty with the ~ marker.

Is there a way to ignore all other boxes in the initial decryption so subsequent decryptions are fully successful, regardless of the state of other boxes in the initial decryption saves?

[kwsch]

The key file is only created if the program is able to dump from the initial six slots it is checking for; after breaking is done, the two save files are copied to their key positions and all further unlocking logic does not happen until breaking is applied again, where the key slots are then adjusted as needed.

Now, here's where I'm unsure. I wasn't the one to add the single-saving feature; what is described above is how I had programmed it.

edit: it appears as if it's not related to the breaking procedure, just compare the key after each new break and see how it is changing. Try and extract data manually too.

[Cu3PO42]

I really don't see how my code could affect this.
What @kwsch described above doesn't happen, the indirection is not big enough. We're not talking references to references, but just references. And since updating a reference won't update the reference it originally copied, this won't screw up.

And I can't reproduce this behavior with any of my games. Could you share the saves in question, please?

[Kaisonic]

Hm... I did the same testing in the original double-save version of KeySAV2, and I get the same behavior. Breaking with either set of "others full" saves results in a single slot not appearing in the "full" save, and appearing with a ~ Pokemon in the "empty" save. Breaking with either set of "others empty" saves results in seeing all 930 Pokemon in the "full" saves, and seeing no Pokemon in the "empty" save.

Here are the saves I'm using (from Y): http://www.kaisonic.net/keysav2testsaves.zip

[Cu3PO42]

Thanks for sharing, @Kaisonic! I can reproduce now, although I have absolutely no idea why this is as of yet.